A scanned PDF is really pictures of pages: there’s no text to select or search. When you open one, Seben shows a banner offering to Run OCR.
Click it and Seben recognizes the text on every scanned page. You can then select and search that text in Seben’s viewer, and search highlights matches on the scanned page.
The recognized text is written into the document itself, invisibly, so the page looks the same. It stays through later changes and comes with the file when you download it, so other PDF readers can search and copy it too.
Once a scan has been through OCR, the AI agent and extraction read its text like any other text. The agent is told which text was recognized from a scan.
Export to Word & Excel recognizes scanned pages on its own, so you don’t need to run OCR first.